At NeurIPS, Michael Hall took the stage in the Exhibit Hall to showcase “Arm AI Acceleration in Action: ExecuTorch + ONNX on Mobile and Edge.” The session walked through how Arm optimizations translate directly into real performance gains for AI workloads, running side-by-side LLM inference and computer vision models on SME2-enabled Arm CPUs and Android devices. Key takeaways: ➡️Arm-optimized ExecuTorch and XNNPack significantly boost performance on mobile and edge devices ➡️A clear optimization pipeline researchers can replicate ➡️Practical, reproducible workflows for running LLMs and CV models efficiently on Arm-based devices A huge thank you to everyone who joined the session. If you missed it, stop by booth #622 to see our demos in action!
Android Edge and deployment of Deep Learning Networks in it https://www.jkuse.com/dltrain/deploy-dl-networks/edge-native-service/j7-app ..very good tutorial..
Is this for training or inference?
thanks for the post. Could you please give us some more links on the AI topic at ARM ? Especially when people are familiar with PyTorch etc...